LiveKit Community

<http://cloud.livekit.io|cloud.livekit.io> is still down for you?

We just migrated datacenters for both <http://docs.livekit.io|docs.livekit.io> and <http://cloud.livekit.io|cloud.livekit.io> — they’re up for me now, please let me know if they’re still down for you and we’ll investigate

Two questions;
1. How did you figure out Egress was available in Cloud? I’m asking so we can make it clearer in our documentation/marketing.
2. Which storage service do you need?
cc <@U01UB0LHTTQ> for visibility 

LiveKit Cloud should not require additional configuration in order to use it. In terms of access keys/etc. You can actually send it in with each Egress request.

<@U01KZ0SCDKL> <@U01UB0LHTTQ>
1. The hint was in a dashboard, where I saw Egress section. But I had to use livekit-cli to confirm.
2. Since I was saving files locally (on server) I was thinking about something like nginx storage server. But then I found localstack.cloud and <http://s3ninja.net|s3ninja.net> for s3 emulation. It may sound strange, but some customers don't want to store their data at gc, azure or aws. So thats why I was asking about some open source solution that can be deployed on any server. Are you looking into something like that in the future?

you can use any s3-compatible storage with egress - some users are using <https://min.io/|minIO>

Thanks. I found it too but didn't look into it yet.I guess I'll stick with this solution for now

hi <@U0527GV4D6Y>
the `startAudio` is only visible in cases where the browser blocks auto-play of audio.
In those cases it is necessary to initiate audio playback with a user gesture (thus the button).
However it’s unexpected that some people wouldn’t be able to hear each other even after clicking the button.
Can you observe a pattern in which users seem to have trouble hearing others? (e.g. browser / extensions etc.)

<@U02EU2ZPTQU> ah i see. so there isn't a pattern to it we've noticed, is there some logging we can grab from the client side for you to check it out when it happens?

is there a possibility maybe the audio tracks of all the other participants hasn't been fully populated for this to work here: <https://github.com/livekit/client-sdk-js/blob/9d4fc819476deef2e6bc90ba21e4b717dbecaefd/src/room/Room.ts#L659>

after audio playback has started once all subsequent `attach` calls will auto play, so it shouldn’t matter if all/which tracks are populated in that array when running looping over it.
you can listen to `RoomEvent.AudioPlaybackStatusChanged` and log that additional on the client side, this would definitely be helpful to see!

<@U02EU2ZPTQU> sorry really struggling to figure out how to add listeners when using `LiveKitRoom` from livekit/components-react, i.e. i'm trying to add a listener for this page: <https://github.com/livekit-examples/meet/blob/main/pages/custom/index.tsx>

happy to help!
this isn’t overly clear in the meet example. What you would want to do is to construct the room manually first and then pass it to `&lt;LiveKitRoom room={myRoom} … /&gt;`.

so something like
```... 
import { Room, LogLevel } from 'livekit-client';

export default function CustomRoomConnection() {
  ...
  const myRoom = new Room();
  const onPlayBackStatusChanged = React.useCallback((status) =&gt; {
    // actually handle the event
  }, []);
  myRoom.on(RoomEvent.AudioPlaybackStatusChanged, onPlaybackStatusChanged)
  ...
  return &lt;LiveKitRoom room={myRoom} .... /&gt;
}```

if you join the room on <http://example.livekit.io|example.livekit.io>, do you see the track?

can you try doing
```pub, err := room.LocalParticipant.PublishTrack(track, &amp;lksdk.TrackPublicationOptions{})```
and using `pub.SID()` instead of `track.ID()` in your request?

One more question related to Egress. Is there any way I can tell egress to split file by files with size or time limit. For example I want to split recorded video track by 10 minutes. Or I have to implement it on my side to stop/start recording every 10 minutes ?

I’ll leave the answering to the experts (<@U01UB0LHTTQ>), but maybe this will be helpful. There is a SegmentedFileOutput: <https://docs.livekit.io/server/egress/#segmentedfileoutput>
where you can set `segment_duration` .

Segmented file will split the files with a time limit into an m3u8 playlist, but if you want to specifically have mp4s you would need a process that stops the egress and starts a new one

Hi <@U04H7PB43H6>, can you explain a bit more about what you’re asking for, pls? Are you asking if we can support i18n for component labels? If so, to which language?

hi <@U04H7PB43H6>
for meet we do not have support for I18N right now.
there’s an issue on the components repo <https://github.com/livekit/components-js/issues/452> that you can chime in on.
we haven’t decided yet on a path on how to support i18n, but it’s definitely interesting!

<@U01KZ0SCDKL> <@U02EU2ZPTQU>

Thank you for the prompt response :slightly_smiling_face:  Yes, I am trying to, in a simplified way, translate the elements of "meet"  into Portuguese.
Thank you for involving me in the discussion :)

<https://docs.livekit.io/getting-started/cli-setup/>
these commands are one line but writing this was makes it look like they are seperate lines:

```livekit-cli create-token \
    --api-key devkey --api-secret secret \
    --join --room my-first-room --identity user1 \
    --valid-for 24h```
```livekit-cli create-token --api-key devkey --api-secret secret --join --room my-first-room --identity user1 --valid-for 24h```

<https://github.com/livekit/livekit-cli/tree/v1.2.5>
doesn't explain clearly what to do after `livekit-cli project add`

same here:

```livekit-cli join-room \
    --url <ws://localhost:7880> \
    --api-key devkey --api-secret secret \
    --room my-first-room --identity bot-user \
    --publish-demo```

```livekit-cli join-room --url <ws://localhost:7880> --api-key devkey --api-secret secret --room my-first-room --identity bot-user --publish-demo```
and here:
<https://github.com/livekit/livekit-cli#usage>

concepts should be mentioned later.<https://docs.livekit.io/concepts/livekit-sfu/>

first let the user experience creating their own simple video sharing asap

There should be examples for different platforms(web: js, react; native: android, ios etc) here: <https://docs.livekit.io/client/connect/>

It is not clear where to put this code or how to get started:
<https://github.com/livekit/server-sdk-js/tree/main/examples/webhooks-http>

hey <@U03GQ47GNF4> — thanks for sharing these. <@U02MC5V3C2H> <@U02LTG3D3K8> are the right folks to chat with about this. <@U01KZ61GV5Z> for visibility.

Hi Oleksandr! Thanks for telling us all this. Responding to your points in order:
1. Better user feedback is definitely something we want to provide. Right now we’re working on the first part of this, which is to make our connection quality score more accurate. Is there a specific thing on the Dashboard that you want to be able to show users?
2. The Overall Quality graph is intended to approximate a <https://en.wikipedia.org/wiki/Mean_opinion_score|Mean Opinion Score> based on track metrics, but as you note it currently is not very accurate. Improving this is on our backlog!
3. Simulcast is always on by default. This is briefly mentioned in the docs but I’ve created a ticket to explain it in more detail. If you don’t specify LOW/MEDIUM/HIGH layers, default layers are computed (<https://github.com/livekit/client-sdk-flutter/blob/2b1e7f2e5f9f0007f604137fe4036a66941ec685/lib/src/utils.dart#L237|here in the flutter sdk source>). If you do specify the layers, it’s a list of VideoParameters from low to high, as in <https://github.com/livekit/client-sdk-flutter/blob/2b1e7f2e5f9f0007f604137fe4036a66941ec685/lib/src/types/video_parameters.dart#L60|the preset definitions in the sdk source>.
4. We are currently building a better search feature to let you search for specific users! It required a bit of a refactor but it’s on the way :smile: 

do you use `adaptiveStream`? If so that's why those layer changes are taking place automatically. The subscriber's client is measuring the UI elements that video tracks are attached to, and updating the SFU automatically

<@U01KZ61GV5Z> no, I don't use `adaptiveStream` . It is set to false. That's why I was confused.

<@U02MC5V3C2H> Thank you for the explanations. I guess I totally messed up layers when I first tried to configure them. I set them in the wrong order :sweat_smile: (from the highest res to the lowest)

Regarding the data I want to show to testers: I would like to show them bitrate and packet loss plots so they can compare their experience during the call with those plots and give us more consistent feedback. I guess I can write my own logger that will save this data to influxDB (for example) but it would be really helpful if this data via some sort of API. That way I wouldn't redo the work that was already done by you.

can you share an example session with us? we'll take a look

Sorry, I don't get it. What do you mean by that? What exactly should I share?

Hey <@U03GQ47GNF4>, I'm referring to a link to the session view in the dashboard where you were seeing multiple `Publishing quality changed due to subscribers requesting HIGH layer` flags.

We'd like to check on our side to understand if they are expected

<@U01KZ61GV5Z> Unfortunately I cannot find the session. I didn't save the url/id at the time I was viewing sessions. I was trying to find the right session but no luck, although I think I've visited the right one and this time there were no "*Publishing quality changed due to subscribers requesting HIGH layer"* marker on the plot.
I'll share it in this thread and mention you if I find one.

<@U01KZ61GV5Z> Hi! I forgot to mention. I set `adaptiveStream`  to `false`  because after some time on Android device (depends on device) the video of other participants turns into slide-show, while their local video is displayed without any issue. This behaviour can be reproduced on multiple android devices even with the example app for Flutter. Do you have any idea why it may happen?

so without adaptiveStream, it behaves correctly for other participants?

by "slide-show", do you mean the framerate drops to &lt;5fps ?

Hi <@U01KZ61GV5Z>. Sorry for a bit late response. Yes and yes (to both your questions). And the strangest thing is that it only happens on Android devices.
I'm taking part in a project where we have some additional UI/UX logic to a call and there this behaviour can be reproduced even with `adaptiveStream=false`.

CC: <@U01NZKTNE1H>.  Are you able to reproduce this with our example app <@U03GQ47GNF4>? What's the best way for us to reproduce this issue?

<@U03GQ47GNF4> do you have an estimate of how long it takes before this happens? I’ll try to repro, but will be helpful to know what timescale I’m looking at here

Hi <@U01NZKTNE1H>. It depends on device. For example, on Xiaomi Mi 9T 6/64GB Carbon Black it became noticeable after 15-20 minutes.

Hmm, just tried on the android sdk, didn’t seem to repro, would it be possible for you to try building the sample app from <https://github.com/livekit/client-sdk-android>? I’ll try with the flutter example app, but curious if there’s a difference there

Hi <@U01NZKTNE1H>. I've tested android sample app and I was able to reproduce. I used 2 android devices: *Umidigi F2* and *OnePlus Nord CE 5g (12GB RAM).*  Video from OnePlus was pretty good on F2 however it had freezes from time to time, while the video from F2 on OnePlus became a slide-show over time. Interesting observation: when I collapsed (during the call) the app on OnePlus the video from F2 got back to normal (like at the very beginning of the call). I had this strange behaviour several times in the app I'm currently working on.

this process is automatic. at the moment there aren't any APIs that returns which server each room is on

is there any document how node selector works

caddy vs caddy-l4, what is different between these two? layer4 is better performance?

we need layer4 routing in order to front the embedded TURN server

I have made sure the ports 7880 &amp; 7881 are open on ufw

```streamRequest := &amp;livekit.RoomCompositeEgressRequest{
		RoomName: roomName,
		Layout:   roomCompositeLayout,
		Output: &amp;livekit.RoomCompositeEgressRequest_Segments{
			Segments: &amp;livekit.SegmentedFileOutput{
				FilenamePrefix: fileNamePrefix,
				PlaylistName:   fileNamePath,
				Output: &amp;livekit.SegmentedFileOutput_S3{
					S3: &amp;livekit.S3Upload{
						AccessKey: h.S3BucketConfig.AccessID,
						Secret:    h.S3BucketConfig.Secret,
						Region:    h.S3BucketConfig.Region,
						Bucket:    h.S3BucketConfig.BucketName,
					},
				},
			},
		},
	}

	return h.LivekitEgressClient.StartRoomCompositeEgress(ctx, streamRequest)```

This looks like a authentication issue, but it is unclear when that is from….

Apology for the late reply (i was asleep).

<@U01KZ61GV5Z>
```liveKitEgressClient := lksdk.NewEgressClient(
   cfg.FlagConfig.LiveKit.Host,
   cfg.FlagConfig.LiveKit.Key,
   cfg.FlagConfig.LiveKit.Secret,
)```


and how is the server hosted? Cloud or OSS?

We are using AWS (ECR and ECS). <@U01KZ61GV5Z>

<@U01KZ61GV5Z>, I marking this issue as resolved. The bug is from our end. Thank you!

ListEgress :smile: <https://docs.livekit.io/server/egress/|https://docs.livekit.io/server/egress/>

I think the disconnect this is..

your `nodejs` server is different from livekit-server. There are three components here:
• livekit client SDK
• livekit server (runs on port 7880)
• your backend server (NodeJS)
Both livekit client SDK and your backend server should point to livekit-server at port 7880

But it is really difficult to understand that from the docs. I setup 100ms and videosdk in 2 hours. Been trying to setup livekit for 2 days. I am almost gonna give up.
Which one is livekit server? Do I have to set it up also?
<https://docs.livekit.io/getting-started/server-setup/> this one?
How do I deploy that?

So the question is, are you looking for a fully hosted solution? if so, you should sign up for <https://cloud.livekit.io/|LiveKit Cloud>.

LiveKit can also be fully hosted by yourself, and <https://docs.livekit.io/getting-started/server-setup/#running-locally|here> are the instructions to install a copy of the server.

I did sign up! I want to use the Cloud. I have a 4000 user base. I want to move from Agora but I just can't understand the docs and I read thousands of docs. I did the stuff you showed on the page I connected to the server but how can I create my own? It is a windows exe file.
<https://docs.livekit.io/getting-started/server-setup/#using-livekit-cloud|Using LiveKit Cloud> it says. But where is the rest of the instructions?

Got it! in that case, you can ignore LiveKit server entirely (those instructions are geared towards someone hosting their open source servers). Sorry for the confusion there. we have an improved onboarding experience coming shortly. CC: <@U02MC5V3C2H>

In the mean time, if you are using Cloud, please follow <https://blog.livekit.io/mini-tutorial-setting-up-a-livekit-cloud-project/|this tutorial> to get started!

Feel free to ping us in <#C048FRL1N2C|cloud> if you have any question or feedback about using Cloud

Thank you for the replies. Just one quick last question.

I have done the stuff on the page you sent.
I have the client SDK ready.
I have the backend server (NodeJS) ready.
On the client side what do I use for the wsurl?  For example here: `await room.connect('<ws://localhost:3000>', 'userToken');`
 or here: `await room.prepareConnection(wssUrl.value);`

Do I use the cloud livekit url? Do I add a port?
<wss://myproject.livekit.cloud>

Let me share my server code just in case:<https://codeshare.io/gLgJqV>

Also, you realllllly need to make the docs open source. There are lot of stuff I want to fix and make more readable. LiveKit is opensource why not the docs?

yeah, the cloud URL `<wss://myproject.livekit.cloud>` would be the server URL that you put into `room.connect`. You do not need to add a port. By default it'll use the standard HTTPS port (443)

yeah! we plan to make the docs open. (as the previous iteration was). We are just doing some clean up right now on it

image.png

Thank you! I finally connected! Thank you for your time!

hi!
with how many parallel screen shares do you start to notice this performance impact?
How many videos are visible in total at the same time ?

generally in order to increase performance on high resolution screens, you could try to change `adaptiveStream: { pixelDensity: 'screen' },` to `adaptiveStream: { pixelDensity: 1 },`

<@U02EU2ZPTQU> around 5 parallel screen shares and 5 just audio tracks with participant placeholders.
thank you. i will try it.

Does the URL look correct to you, or is the wrong format?

You might want to try setting ForcePathStyle = true in the S3Upload object of the request

Many thanks, setting `ForcePathStyle = true` was helpful

try keda : <https://keda.sh/docs/2.10/scalers/metrics-api/>
Haven't tried it though with custom metrics.

or you can use k8s - HPA. You may need think about a Kubernetes Nodes Autoscaler stratergy as it may take 1-3 mins to up the Worker Node VM

This would require some modifications to the code to keep it running at 0 resources. It would need to intercept the start egress request, send a request to scale up, and the node would need to deploy almost instantaneously

What versions are you running? I received another report of this last week which I could not reproduce on the latest

<@U01UB0LHTTQ>
Some additional information:
In the logs I can see that livekit only sends `engress_end` *after* `room_finished` webhook event has been handled. So until `room_finished` event isn’t processed it doesn’t even try to send `egress_end` .

Also I can in the logs that `track egress` participants only leaves the room after the room finished and not after the connected participant left.

```2023-05-15T17:15:36.307Z INFO livekit service/roommanager.go:113 deleting room state {"room": "8f5vbht4ilwlqaqq"}

2023-05-15T17:15:36.309Z INFO livekit service/roommanager.go:455 room closed {"room": "8f5vbht4ilwlqaqq", "roomID": "RM_aqCmzcvr6x92"}

2023-05-15T17:15:36.322Z DEBUG livekit service/roommanager.go:493 RTC session finishing {"room": "8f5vbht4ilwlqaqq", "roomID": "RM_aqCmzcvr6x92", "participant": "EG_Z3aMaNp4QFZG", "pID": "PA_XfjCbz7J2ni6", "remote": false}

2023-05-15T17:15:36.349Z DEBUG livekit service/roommanager.go:493 RTC session finishing {"room": "8f5vbht4ilwlqaqq", "roomID": "RM_aqCmzcvr6x92", "participant": "EG_iifhfMWtYJsd", "pID": "PA_Y5GcuYEdbTCT", "remote": false}

2023/05/15 17:15:36 [DEBUG] POST XXXX/webhook

2023-05-15T17:15:36.693Z INFO livekit service/ioinfo.go:78 egress ended {"egressID": "EG_iifhfMWtYJsd"}

2023-05-15T17:15:36.813Z INFO livekit service/ioinfo.go:78 egress ended {"egressID": "EG_Z3aMaNp4QFZG"}

2023/05/15 17:16:05 [DEBUG] POST XXXX/webhook (status: 504): retrying in 1s (4 left)
.....
2023/05/15 17:17:39 [DEBUG] POST XXXX/webhook (status: 504): retrying in 8s (1 left)

2023-05-15T17:18:16.451Z WARN livekit webhook/url_notifier.go:70 failed to send webhook {"url": "XXXX/webhook", "event": "room_finished", "error": "POST XXXX/webhook giving up after 5 attempt(s)"}

2023/05/15 17:18:16 [DEBUG] POST XXXX/webhook

2023-05-15T17:18:16.599Z INFO livekit webhook/url_notifier.go:73 sent webhook {"url": "XXXX/webhook", "event": "egress_ended"}

2023/05/15 17:18:16 [DEBUG] POST XXXX/webhook

2023-05-15T17:18:16.705Z INFO livekit webhook/url_notifier.go:73 sent webhook {"url": "XXXX/webhook", "event": "egress_ended"}```

I’ve added tests that confirm that if the room continues, the track egress participants leave and the `egress_ended` event is sent without waiting for the room to close.

If the egress and the room end at the same time, we don’t make any guarantees about the ordering of the events. The egress service needs to do things like wait for any dropped packets, close out the pipeline, and upload files, which might take a bit of extra time

Here is the stack trace from another project that seems to reach an endpoint in the stack:
`Invoking error handler due to`
`InvalidStateError: An attempt was made to use an object that is not, or is no longer, usable`
`getStats@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/Build/iSeeWebGL.framework.js.jgz:9:109293`
`poll@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:77:21`
`setTimeout handler*poll@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:75:18`
`setTimeout handler*poll@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:75:18`
`setTimeout handler*poll@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:75:18`
`setTimeout handler*poll@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:75:18`
`setTimeout handler*construct@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/index.html:73:14`
`PCTransport@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/Build/iSeeWebGL.framework.js.jgz:9:334470`
`configure@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/Build/iSeeWebGL.framework.js.jgz:9:343634`
`join@https://public-isee2020dev.s3.ap-southeast-2.amazonaws.com/Downloads/Client/Web/Build/iSeeWebGL.framework.js.jgz:9:340875`

<@U022KB72NFJ> Further testing on Firefox using multiple *desktop* platforms has revealed that the above situation only occurs on Windows. Firefox on Linux seems to be fine. However a Firefox client running on MacOS will cause this exception on a remote client when the Mac client connects (remotes were using Chrome and Edge):

Which SDK &amp; version are you using? and Is it Cloud or self-hosted server (if so which version) that you are connecting to?

flutter SDK livekit_client: ^1.2.1. it is VM hosting with docker
On a AWS version is 1.3.5

if I were to guess, the server likely didn't have a TURN server configured.. and/or there are problems with firewall setup. give <https://livekit.io/connection-test|this> a try

IMG_20230517_114200.png

IMG_20230517_114449.png

IMG_20230517_114302.png

IMG_20230517_114411.png

<@U01KZ61GV5Z> this is my testing 2 to 3 times with my link and tokens

this seems to indicate some firewalls/routing isn't configured properly..

<@U01KZ61GV5Z> Can you elaborate on why you think this is firewall . I would expect it to fail everytime if it was firewall issue ? Do you know what port maybe blocked

AWS would block all ports by default.. how is the server configured? and are the ports listed on <https://docs.livekit.io/oss/deployment/ports-firewall/|this page> open?

I meant the host system, I have never seen this before but it seems like an emulation problem

hmm I'm not aware of a way to do this today. we would like to make the load tester better though.

There was an incompatible protocol change for ingress during the past release. Please make sure to run the latest version of livekit-server, ingress, and the cli.

If you still have problems after that, please provide all the versions of the different pieces.

IMG_2444.jpg

IMG_2445.jpg

IMG_2446.jpg

Which SDK are you using? and the version?

Are you using Cloud ? Or is this a self-hosted server?

1) ruby sdk и react sdk  <https://github.com/livekit/server-sdk-ruby|https://github.com/livekit/server-sdk-ruby> и <https://github.com/livekit/components-js|https://github.com/livekit/components-js>

2) self-hosted server

I will be very grateful for your help. Sometimes we also catch a case when a microphone and a camera are not used when entering a webinar - there are no disconnects, and if there is at least something, then disconnects begin

"disconnects" is tricky to debug, because WebRTC servers are difficult to set up correctly. I would suggest testing against LiveKit Cloud first. If you also see disconnect there, then it means there's something going on with the client integration or ISP/firewalls.

Which exception are you looking at? Is this about an SDK? (if so, which one?)

I want an exception for nodejs in livekit server sdk,
I also want to see the error code and message specification of the livekit server

I see, if I understand correctly, you are trying to anticipate failure modes while using server APIs?

We currently don't have a listing of all of the possible status codes, though they should follow general HTTP guidelines

yes and i also find status code list in here <https://github.com/livekit/psrpc/blob/main/errors.go> thank you for replying

IMG_1268.png

IMG_1269.png

Hey <@U04TVCSDSHG>, this is most likely due to how GStreamer publishing the stream. Once the packets reach LiveKit, there is no buffering on our side (the packet spends &lt;1ms within LiveKit SFU before getting forwarded on). The reason why it's delayed is either
• GStreamer isn't able to keep up with real-time publishing. so it adds a bit of skew overtime. (you can take a look at queues/etc that are in your pipeline)
• There's something wrong with the PTS that your pipeline is generating, that's falling behind just slightly.

Thanks for the response! Do you know why just restarting LiveKit resolves the issue, if it is a GStreamer problem? I would think that if it were a GStreamer issue, restarting LiveKit wouldn't do anything, and restarting the gst pipeline would

When you restart LiveKit, the track would have to be republished again, effectively "resetting" what the browser thinks is realtime.

How do you handle the server going away on the client side? I imagine you are publishing with Go SDK?

sorry, I should have clarified that we're rejoining the room, not restarting the entire server, but yeah we're on the Go SDK 1.4.2

does the track republish then as well? I would guess so but just want to double check

rejoining the room as the publisher? or rejoining the room as viewers?

ok, that's a bit difficult to know.. does it correct itself if the viewer joins?

we haven't actually tested that, I can try that tomorrow and get back to you

if resetting the viewer gets you to live. then it means it's a timestamp/fps issue. with the Go SDK, FPS would have to match the input exactly.

checked this morning, and the latency is still present when a new viewer joins the room

ok, that means the issue is on the publisher side.

Screenshot 2023-05-17 140421.png

Hi Satyadeep, please keep posts in a single channel and avoid duplication :pray:

For this issue, are you terminating TLS connections?

terminating TLS connections can you please explain a bit on this

all I am doing here is getting hits for livekit and passing them to livekit

My command to start livekit-server runs like : l`ivekit-server --bind 0.0.0.0 --dev --keys devkey: secret`

<https://docs.nginx.com/nginx/admin-guide/security-controls/terminating-ssl-http/>

yes, the same, so what do you suggest , should we move to nginx plus , or any other config param I can use

I would like to take a moment to rephrase what I am trying to achieve:
I am running livekit-server on a ubuntu18.04  machine , and am trying to pass on any request for it via reverse-proxy in nginx , for endpoint `livekit`

Hi <@U01KZ61GV5Z>, I am not sure if this is too much to ask, but can we get on a huddle, at any time of your convenience , and may be you can guide me on what is missing ? Thanks

sorry I cannot help with this. setting up TLS is not something that we can help with. There should be sufficient tutorials online since that isn't LiveKit specific.

ok , however what I want to understand is

Hi! I'm trying to enable my camera and microphone by using `await room.localParticipant?.enableCameraAndMicrophone();`
but my pc has no plugged in camera. for this reason it gives me the error. but the problem is, If it unable get the camera, then `RoomEvent.LocalTrackPublished` and `RoomEvent.TrackSubscribed` those events aren't work. In this two events I'm rendering the participant's streaming video.

This feature is indeed in our roadmap, but we don’t have time yet, this is a feature similar to insertable streams, the related issue is here,
<https://github.com/livekit/client-sdk-flutter/issues/201>

Your research enlighten me that perhaps there is a simpler alternative to implementing video filters

Which part in particular is poor? The latency? Accuracy? One trade-off we made is optimizing for latency over accuracy. We send smaller audio context windows (I can’t remember exactly, but less than the recommended 100ms) to Google. The goal with KITT is to get as close to real, human-like interaction as possible. Depending on your needs, you may want to tweak the context window being sent to STT. We also may have some other configuration parameters set on the Google Cloud side which affect performance (latency or accuracy) in some way. <@U022KB72NFJ> knows more than I do here.

from the guideline, the audio window is 20ms as I remembered